Update documentation

Author: lorenzotomada

README.md

@@ -94,10 +94,12 @@ Once the needed modules are loaded, you can regularly install via `pip` using th
 python -m pip install .
 ```
 
+Notice that, under the hood, a couple of complex things are going on: the `C++` functions are compiled to a shared library using `pybind11` (automatically taking care of the dependence on `Eigen`), then the library is placed in the right folder and suitable paths are updated, so that the `Python` code is automatically able to import `C++` utils.
+
 ## Additional note
 There is the possibility that you may need to write the following command:
 ```bash
 conda install -c conda-forge gxx
 ```
 Notice that you should first try *without* using that command, as the normal installation is expected to work.
-This is likely a consequence of the fact that we are using a `setup.py` file in order to automatically compile a `C++` file when `pip install .` is executed.
+This is likely a consequence of the fact that we are using a `setup.py` file in order to automatically compile a `C++` file when `pip install .` is executed.

docs/Documentation.ipynb

@@ -27,3 +27,10 @@ Submodule: helpers_secular
    :undoc-members:
    :show-inheritance:
 
+Submodule: parallel_tridiag_eigen
+================
+
+.. automodule:: pyclassify.parallel_tridiag_eigen
+   :members:
+   :undoc-members:
+   :show-inheritance:

docs/pyclassify.rst

@@ -10,7 +10,7 @@
 from pyclassify.eigenvalues import Lanczos_PRO
 
 
-seed = 84
+seed = 10
 np.random.seed(seed)
 
 

scripts/mpi_running.py

@@ -20,7 +20,7 @@
 from mpi4py import MPI
 
 
-seed = 84
+seed = 10
 np.random.seed(seed)
 
 
@@ -111,7 +111,8 @@
 delta_t_parallel = time_after_parallel - time_before_parallel
 
 total_mem_children = comm.reduce(delta_mem, op=MPI.SUM, root=0)
-total_time_children = comm.reduce(delta_t_parallel, op=MPI.SUM, root=0)
+# total_time_children = comm.reduce(delta_t_parallel, op=MPI.SUM, root=0) #only if we want the sum!
+total_time_children = delta_t_parallel
 
 # Collect the information across all ranks
 if rank == 0:

scripts/profiling_memory_and_time.py

@@ -32,7 +32,7 @@ conda init
 conda activate devtools_scicomp
 
 # Ranges over which we iterate
-n_processes=(1 2 4 8)
+n_processes=(2 4 8)
 matrix_sizes=(10 50 100 500 1000)
 
 last_dim="${matrix_sizes[-1]}"

shell/submit.sbatch

@@ -1,7 +1,7 @@
 #!/bin/bash
 
 # Ranges over which we iterate
-n_processes=(1 2 4 8)
+n_processes=(2 4 8)
 matrix_sizes=(10 50 100 500 1000 1500)
 
 last_dim="${matrix_sizes[-1]}"